Self-Aligned Bipolar Transistor Process

This application is related to my commonly assigned co-pending patent application entitled, "Bipolar Process Using Selective Silicon Deposition," by Pfiester, Ser. No. 07/459,845, filed Jan. 2, 1990.

The present invention is related to semiconductor device fabrication in general, and more specifically, to fabrication of semiconductor devices having bipolar transistors.

Semiconductor manufacturers are continually trying to reduce the size of individual devices within an integrated circuit in order to reduce the overall integrated circuit size. However in doing so, fabrication of the devices becomes more difficult, particularly for complex device types such as bipolar, BiMOS, and BiCMOS devices. In an effort to simplify fabrication of small, complex devices, manufacturers are attempting to incorporate as many self-aligned processing steps as possible in order to reduce the number of masking operations. In the fabrication of semiconductor devices having bipolar transistors, many self-aligned fabrication processes have been demonstrated, yet most bipolar devices formed from these processes also have undesirable levels of base and emitter resistance. Therefore, a need exists for an improved process for forming a bipolar transistor in a semiconductor device. More specifically, a need exists for a fully self-aligned process which forms a bipolar transistor having low base and emitter resistance.

The previously mentioned needs are achieved, along with other advantages, with the present invention. In accordance with the present invention, a method for forming a self-aligned bipolar transistor in a semiconductor device is provided. In one form, a semiconductor substrate having an active device region is provided. A patterned masking layer having an opening therein is formed overlying the active device region. A sidewall spacer is formed within the opening of the patterned masking layer. The active device region is selectively doped to form a first base region, using the sidewall spacer and the patterned masking layer as a doping mask. An emitter electrode is formed on the first base region by selective deposition, and a second base region is formed in the active device region by doping. Metal silicide contacts to each of the second base region and emitter electrode are formed and the sidewall spacer is removed.

FIGS. 1-7 are a series of cross-sectional views illustrating the formation of a bipolar transistor in a semiconductor device, in accordance with the present invention.

Bipolar transistors in semiconductor devices take one of two conductivity forms, either NPN transistors or PNP transistors. While the description and figures included in this application discuss formation of an NPN transistor, the present invention can be used to form PNP transistors by modifying conductivity types. Likewise, the present invention can also be used in fabricating bipolar transistors in BiMOS and BiCMOS devices. Although bipolar devices can be fabricated in many ways and may appear to be structurally different from one another, bipolar devices always have three common elements, namely an emitter, a collector, and a base. In an NPN bipolar transistor, the emitter and collector have an N-type conductivity while the base has a P-type conductivity. In a PNP transistor, the conductivities for the elements is reversed.

In forming NPN bipolar transistors in an integrated circuit, some practices and fabrication techniques have become quite common. For instance, most integrated circuits are formed on a P-type silicon substrate. To form bipolar devices within this silicon substrate, an N-type epitaxial silicon layer is grown over the P-type substrate. Between a portion of the epitaxial layer and the substrate, a heavily doped N-type impurity is introduced to form an N.sup.+ buried layer which functions as a collector of the transistor. Although the method in which the collector is formed is not standardized in the semiconductor industry, the use of a buried layer as the collector is quite common. The remaining steps in forming an NPN transistor differ significantly from manufacturer to manufacturer. FIGS. 1-7 illustrate one example of the formation of a bipolar transistor in accordance with the present invention. It is important to realize that modifications in the fabrication process described and illustrated herein can be made without departing from the scope of the invention.

FIG. 1 illustrates, in cross-section, an initial structure of a semiconductor device from which a self-aligned bipolar transistor is formed, in accordance with the present invention. A substrate 12 is a semiconductor of N-type conductivity for forming an NPN transistor, and P-type for forming a PNP transistor. Substrate 12 is usually a single crystal material of silicon or an epitaxial layer formed on another underlying substrate (not shown). Below substrate 12 may be a heavily doped buried layer (not shown) which is a collector of the bipolar transistor having opposite conductivity from substrate 12. Furthermore, a collector contact (not shown) may be included to electrically access the buried collector.

Also illustrated in FIG. 1 are isolation regions 14 which are typically made of a thick oxide material and are formed on substrate 12 by conventional methods. Other isolation regions can also be included in the device. Isolation regions are used to electrically isolate adjacent devices from one another, although FIGS. 1-7 illustrate only one device. Between isolation regions 14 and within substrate 12 is an active region 13. It is within active region 13 that the bipolar transistor is formed. Overlying active region 13 is an oxide layer 16 which, in one form of the present invention, serves as an etch stop, prevents damage to substrate 12 during doping, and isolates emitter and base contacts. Each of these purposes will be illustrated or described at a later point. A patterned masking layer 18 is formed overlying substrate 12 such that an opening 19 having walls is formed over active region 13. Materials used for masking layer 18 should be selected based on etch selectivity requirements which will be mentioned throughout this description. For instance, one requirement of masking layer 18 is that it can be etched selectively to isolation regions 14 and oxide layer 16. One suitable material for masking layer 18 is Si.sub.3 N.sub.4.

A sidewall spacer 20 is formed on the walls of opening 19, as illustrated in FIG. 2. Although illustrated as two sidewall spacers, one continuous sidewall spacer, which appears as two spacers when viewed in cross-section, may be formed in accordance with the present invention. However, it should be understood that the present invention may be practiced using two separate sidewall spacers. Materials for use in forming sidewall spacer 20 are those which can be etched selectively to masking layer 18 and underlying oxide layer 16. Likewise, masking layer 18 must have the ability to be etched without etching sidewall spacer 20. For example with a Si.sub.3 N.sub.4 masking layer, TiN would be an appropriate material from which to form sidewall spacer 20. An intrinsic base 22 is formed within active region 13, as illustrated in FIG. 2. Intrinsic base 22 has a conductivity opposite from the conductivity of substrate 12. Intrinsic base 22 is formed by doping the device, for instance, with a P-type impurity such as boron in the case of an NPN transistor. Doping can be accomplished using ion implantation or another known doping procedure. Masking layer 18 and sidewall spacer 20 act as a doping mask during formation of the intrinsic base.

A portion of oxide layer 16 which lies above intrinsic base 22 may remain over active device region 13 during doping to prevent crystalline damage during ion implantation. However, this portion of oxide layer 16 must be removed either prior to or following formation of the intrinsic base in order to form an emitter electrode 24 as illustrated in FIG. 3. Oxide layer 16 can be removed by reactive ion etching or wet etching without damage to other portions of the device. Once a portion of oxide layer 16 is removed, emitter electrode 24 is formed by selective deposition of, for instance, amorphous, monocrystalline, or polycrystalline silicon. Selective deposition is a process of chemical vapor deposition in which the reaction is adjucted so that a material is deposited only on a prepared nucleation site. In the case of selective deposition of silicon, exposed portions of polycrystalline or monocrystalline silicon act as the nucleation site for deposition. With continued deposition, the selectively deposited silicon itself acts as a nucleating site so that additional silicon deposits on earlier deposits to extend silicon upwardly within sidewall spacer 20. The selectively deposited emitter electrode is self-aligned because it deposits only on the exposed intrinsic base 22. While selective deposition of silicon is more common than selective deposition of tungsten, aluminum, platinum, or the like, these materials can also be used in accordance with the present invention.

In order to form an emitter 26, as in FIG. 4, selectively deposited emitter electrode 24 may be insitu doped with an N-type impurity if substrate 12 is N-type or with a P-type impurity if substrate 12 is P-type. Emitter 26 is then formed by a diffusion of the impurities from emitter electrode 24 into the intrinsic base 22. Rather than insitu doping, emitter 26 may be formed by ion implantation or other doping technique following formation of intrinsic base 22 and prior to selectively depositing emitter electrode 24. Furthermore, intrinsic base 22 and/or emitter 26 can be formed following formation of emitter electrode 24. For instance, an undoped emitter electrode may be selectively deposited and then heavily doped to form intrinsic base 22 and/or emitter 26.

Once emitter 26 and emitter electrode 24 are formed, masking layer 18 is removed from the device, as also illustrated in FIG. 4. Removal of masking layer 18 is accomplished using conventional etching procedures, without substantially etching sidewall spacer 20, oxide layer 16, or isolation regions 14. For the case in which masking layer 18 is Si.sub.3 N.sub.4 and sidewall spacer 20 is TiN, a plasma etch mixture of CHF.sub.3, CF.sub.4, and argon or a wet etch using heated phosphoric acid will achieve the desired result. Following removal of masking layer 18, the device is doped again to form an extrinsic base 28 which is of the same conductivity as intrinsic base 22. Extrinsic base 28 can be formed, for example, by ion implantation. Just as sidewall spacer 20 appears as two individual spacers, extrinsic base 28 may be one base which when viewed in cross-section appears as two or it may indeed be two extrinsic base regions. Also, as in the formation of intrinsic base 22, oxide layer 16 may remain above active device region 13 during doping to reduce crystalline damage, however must be removed at a point in the processing to expose extrinsic base 28. Removal of oxide layer 16 is accomplished, for example, by reactive ion etching.

After exposing extrinsic base 28, a refractory metal layer 30, such as titanium or cobalt, is deposited over the device as illustrated in FIG. 5. Notice that refractory metal layer 30 is not only in direct contact with extrinsic base 28, but also with emitter electrode 24. This condition enables silicide contacts to be formed on the extrinsic base 28 and on emitter electrode 24. FIG. 6 illustrates these contacts, namely a silicide base contact 32 and a silicide emitter contact 34. Silicidation is a process in which a refractory metal in contact with silicon undergoes a metallurgical reaction, particularly during a thermal cycle, to form a refractory metal of silicide. A refractory metal silicide is a compound containing both refractory metal and silicon atoms. An example of a known silicidation process is to react titanium with silicon to form TiSi.sub.2. With respect to FIG. 5, silicidatin will only occur on those portions of the device in which the refractory metal layer 30 is in contact with substrate 12 and emitter electrode 24. Thus, silicide contacts are formed on extrinsic base 28 which is within the substrate and on emitter electrode 24. Because the contact formation does not involve any masking or lithography steps, it is self-aligned silicide process, also referred to as a salicide process.

Those portions of refractory metal layer 30 which do not contact silicon or polysilicon do not undergo a metallurgical reaction, for example those portions which are in contact with sidewall spacer 20 and isolation regions 14. The unreacted portions are removed, as illustrated in FIG. 6, while reacted portions, silicide base contact 32 and silicide emitter contact 34, remain substantially intact. It is advantageous to also remove sidewall spacer 20 simultaneously with the removal of unreacted portions of refractory metal layer 30. For instance, a TiN sidewall spacer can be removed easily during the removal of unreacted titanium by using a heated hydrogen peroxide solution or a hydrogen peroxide and ammonia hydroxide solution. If sidewall spacer 20 cannot be removed at the time the unreacted refractory metal is removed, an additional processing step to do so is required.

A remaining step is to lightly dope the device to form a link region 36 which adjoins extrinsic base 28 with intrinsic base 22. Link region 36 is illustrated in FIG. 7, and is of the same conductivity type as the extrinsic and intrinsic bases. No masking is required in forming link region 36. Other portions of the device will also be doped. However, the doping concentration required to form link region 36 is much lower than the doping concentrations of other portions of the device, and therefore will have no impact on performance of the device. Because the link region is formed as one of the last steps in forming a bipolar transistor in accordance with the present invention, the link region can remain shallow. Many bipolar processes perform a link implant earlier and thus have a disadvantage in that the link region diffuses, in other words becomes deeper, during thermal cycles of subsequent processing steps. Shallow link regions, such as link region 36 in FIG. 7, are desirable to reduce series resistance in the base regions.

The present invention has several advantages over known bipolar transistor fabrication processes. One advantage is that a bipolar transistor formed in accordance with the present invention has lower base and emitter resistance than other bipolar transistors. Because the extrinsic base and emitter contacts are formed of silicide, base and emitter resistance is improved. In addition, the present invention provides for a fully self-aligned structure which enables the emitter and extrinsic base regions to be in very close proximity to one another while maintaining a large surface area, thus further reducing resistance. The self-aligned bipolar transistor taught herein not only has reduced resistance, but also simplifies the fabrication process particularly for very small devices. Another advantage, as discussed earlier, is that a link region is formed relatively late in the fabrication process and therefore remains shallow.

Thus, it is apparent that there has been provided, in accordance with the present invention, a process for forming a self-aligned bipolar transistor in a semiconductor device that fully meets the advantages set forth previously. Although the invention has been described and illustrated with reference to specific embodiments, it is not intended that the invention be limited to these illustrative embodiments. Those skilled in the art will recognize that modifications and variations can be made without departing from the spirit of the invention. For example, use of the invention is not limited to the materials, etching techniques, doping techniques, or deposition techniques specifically described or illustrated. Any material or process which achieves the desired result of a particular element in accordance with the present invention is suitable and within the scope of the invention. Furthermore, the described process can be used to form PNP or NPN transistors or both, either alone or in combination with N-channel, P-channel, or CMOS transistors. While the description has emphasized the used of a silicon substrate and selective deposition of either amorphous, monocrystalline, or polycrystalline silicon, the invention is not limited to use of these materials. Therefore, it is intended that this invention encompass all such variations and modifications as fall within the scope of the appended claims.